Yuille & Cutshall

To record and evaluate witness accounts considering accuracy and the kind of error made

To examine issues raised by laboratory research

21 witnesses observed a shooting where a man was killed and another seriously injured. They were initially questioned by the police describing the events and answering questions.

4/5 months after the event Yuille & Cutshall tracked down the witnesses and asked if they could interview them. 13 agreed to participate.

Incorporated in the questions were 2 misleading questions:

Half were asked if they saw A broken headlight whilst the other half were asked if they saw THE broken headlight. (there was no borken headlight really)

Half were asked if they saw THE yellow quarter panel and the other half A yellow quarter panel (the panel was really blue)

They also assessed participants level of stress at the time and if they had any subsequent problems like sleeplessness

A careful scoring procedure was used so comparisons could be made where qualitative accounts were turned into quantitative data. They divided information into action details and description details

Misleading questions had little or no effect on participants answers as 10/13 said there was no broken headlight

They were highly accurate in their accounts, with little or no change in amount or accuracy over 5 months

5 witnesses closest to the offender reported the highest level of stress at the time and had difficulty in sleeping after the incident

The results differ from the pattern of many laboratory research and show a need for more field experiments to evaluate the generalisability of lab research. It is suggested that Eye-witness testimony is reliable

Eye-witness testimonies should still be acknowledged as useful in court

Stress doesn't necessarily affect memonry and misleading questions do not alter recall either